Optimization of maintenance performance measurement of critical machines in Tea Processing: A case study at Litein Tea Factory

Abstract

In industrial processing, functional and efficient plant equipment is crucial for optimum output. Therefore, proper maintenance of plant equipment can help minimize operational expenditures arising from breakdowns. The challenge has been to establish the correlation of maintenance and manufacturing performance, and the overall operational performances in product quality, product cost and plant availability. Therefore, the main objective of this study was to optimize maintenance performance measurements of critical machines in tea processing, with a case study of Litein Tea Factory, Kenya. Its specific objectives were to: identify a critical equipment in tea processing plant; evaluate a maintenance model for a critical equipment in tea processing plant and optimize the maintenance model of a critical equipment in tea industry. To identify the critical equipment, data were collected using questionnaires and analysed using Statistical Package for the Social Sciences software to evaluate criticality. From the failure mode effects analysis, Crush, Tear and Curl, with an Index of 242, was established as the most critical unit in tea processing. Data on downtime, throughput, operating time, number of failures, failure type and service time were then collected from the Crush, Tear and Curl in Excel sheet and transferred to Minitab worksheet. Probability plot for the parameters in a sample size of 28 was of normal distribution and with P-values of 0.005. Mean and Standard deviations were also tested. Correlations between the dependent (Y) (throughput and number of failures) and independent (X) variables generated a R2 values of 89.06% for throughput and 51.72% for number of failures models. The evaluated models were validated by use of sensitivity analysis to assess how changes in input parameters affect the model output, simulated and the summary statistics derived from Monte Carlo Simulation. Initial process performances were 0.0501 for number of failures and -0.0291 for throughput regressions. Meanwhile, percentages out of specifications corresponded to highs of 59.64% for number of failures and 83.06 for throughput models. Parameter optimization was then undertaken to generate best fit and optimal variables. The results indicated process performance of 2.98 with corresponding 0.00% out of specs for the number of failure regression and 1.17 process performance and respective 0.05% out of specs performance for throughput regression model. In conclusion, utilizing optimized parameters can enable factories to improve on machine availability, reliability, maintainability, and overall efficiency. Future research should sample from many tea factories to help validate the study results. Future research should also endeavour to raise the value of R2 in the regression for number of failures to bring the statistical figures close to the fit line.